AUTHOR=Alim-Marvasti Ali , Pérez-García Fernando , Dahele Karan , Romagnoli Gloria , Diehl Beate , Sparks Rachel , Ourselin Sebastien , Clarkson Matthew J. , Duncan John S. 

TITLE=Machine Learning for Localizing Epileptogenic-Zone in the Temporal Lobe: Quantifying the Value of Multimodal Clinical-Semiology and Imaging Concordance

JOURNAL=Frontiers in Digital Health

VOLUME=Volume 3 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/digital-health/articles/10.3389/fdgth.2021.559103

DOI=10.3389/fdgth.2021.559103

ISSN=2673-253X

ABSTRACT=Background

Epilepsy affects 50 million people worldwide and a third are refractory to medication. If a discrete cerebral focus or network can be identified, neurosurgical resection can be curative. Most excisions are in the temporal-lobe, and are more likely to result in seizure-freedom than extra-temporal resections. However, less than half of patients undergoing surgery actually become entirely-seizure-free. Localising the epileptogenic-zone and individualised outcome predictions are difficult, requiring detailed evaluations at specialist centres. 

Methods

We used bespoke natural language processing to text-mine 3,800 electronic health records, from 309 epilepsy surgery patients, evaluated over a decade, of whom 126 remained entirely-seizure-free. We investigated the diagnostic performances of machine learning models using set-of-semiology (SoS) with and without hippocampal sclerosis (HS) on MRI as features, using STARD criteria.

Findings

Gradient Boosted (GB) decision trees and Support Vector Classifiers (SVC) were the best performing algorithms for temporal-lobe epileptogenic zone localisation (cross-validated Matthews correlation coefficient (MCC) SVC 0·73 ± 0·25, balanced accuracy 0·81 ± 0·14, AUC 0·95 ± 0·05). Models using only seizure semiology did not always improve scores above internal benchmarks. The combination of multimodal features, however, enhanced performance metrics including MCC and normalised mutual information (NMI) compared to either alone (p<0·0001). This combination of semiology and HS on MRI increased both cross-validated MCC and NMI by over 25% (SVC SoS: 0·35 ± 0.28 vs SVC SoS+HS: 0·61 ± 0.27). 

Interpretation

Machine learning models using only the set of seizure semiology (SoS) cannot unequivocally perform better than benchmarks. However, when combined with an imaging feature (HS), SoS enhances epileptogenic lobe localisation. We quantify the value of combining imaging and clinical features in temporal epileptogenic-zone localisation. Despite good performance in localisation, no model was able to predict seizure-freedom better than benchmarks. The methods used are widely applicable, and the performance enhancements by combining clinical, imaging and neurophysiological features can be similarly quantified. Multicentre studies are required to confirm generalisability.

Funding

Wellcome/EPSRC Centre for Interventional and Surgical Sciences (WEISS) (203145Z/16/Z).